shared_skill

Overview

This skill bundles shared Sanskrit reference materials and lightweight Python utilities for script conversion, morphology, sandhi, meter detection, and term lookup. It provides a compact Devanagari guide, multiple transliteration systems, a 150+ term glossary, curated online resources, and ready-to-run scripts used across Sanskrit tools. Use it as a common library when building or integrating Sanskrit processing features.

How this skill works

The package exposes reference files (Devanagari, transliteration tables, terminology, curated online links) alongside Python scripts in shared/scripts. Scripts leverage established libraries (indic-transliteration, sanskrit_parser, vidyut) to perform transliteration, sandhi join/split, dhatu lookup, and chandas (meter) detection. Installation uses a simple dependency sync command and scripts accept clear CLI arguments for quick automation.

When to use it

• Converting text between Devanagari and IAST/Harvard-Kyoto/ITRANS/SLP1/Velthuis
• Splitting or joining sandhi and testing morphological analyses
• Looking up verbal roots (dhatu) or checking classical terminology
• Detecting meter for verses and validating metrical patterns
• Providing a shared resource bundle for multiple Sanskrit tools or integrations

Best practices

• Keep the reference files read-only and import them for consistent terminology across projects
• Use indic-transliteration for bulk conversions and the transliterate.py wrapper for quick CLI tasks
• Run sanskrit_parser or vidyut for production sandhi/morphology tasks; use the simpler scripts for prototyping
• Pin library versions in your environment to avoid breaking API changes
• Prefer Unicode Devanagari input for accuracy; validate outputs against the transliteration table when needed

Example use cases

• Automate conversion of a Devanagari text corpus to IAST before text analysis
• Integrate sandhi splitting into a verse editor to show underlying word boundaries
• Create a teaching tool that highlights Devanagari letters and their IAST equivalents
• Build a poetry validator that flags meter mismatches using chandas.py
• Serve terminology popups in a reader app using the 150+ term glossary